def test_ImageStimulus():
# Create a dummy image:
fname = 'test.png'
shape = (25, 37)
ndarray = np.random.rand(*shape)
imsave(fname, (255 * ndarray).astype(np.uint8))
# Make sure ImageStimulus loaded is identical to dummy image:
stim = ImageStimulus(fname)
npt.assert_equal(stim.shape, (np.prod(shape), 1))
npt.assert_almost_equal(stim.data, ndarray.reshape((-1, 1)), decimal=2)
npt.assert_equal(stim.metadata['source'], fname)
npt.assert_equal(stim.metadata['source_shape'], shape)
npt.assert_equal(stim.time, None)
npt.assert_equal(stim.electrodes, np.arange(np.prod(shape)))
os.remove(fname)
# Resize the dummy image:
ndarray = np.ones(shape)
imsave(fname, (255 * ndarray).astype(np.uint8))
resize = (12, 18)
stim = ImageStimulus(fname, resize=resize)
npt.assert_equal(stim.shape, (np.prod(resize), 1))
npt.assert_almost_equal(stim.data,
np.ones((np.prod(resize), 1)),
decimal=2)
npt.assert_equal(stim.metadata['source'], fname)
npt.assert_equal(stim.metadata['source_shape'], shape)
os.remove(fname)
def test_ProsthesisSystem_reshape_stim(rot, gtype, n_frames):
implant = ProsthesisSystem(ElectrodeGrid((10, 10), 30, rot=rot,
type=gtype))
# Smoke test the automatic reshaping:
n_px = 21
implant.stim = ImageStimulus(np.ones((n_px, n_px, n_frames)).squeeze())
npt.assert_equal(implant.stim.data.shape, (implant.n_electrodes, 1))
npt.assert_equal(implant.stim.time, None)
implant.stim = VideoStimulus(np.ones((n_px, n_px, 3 * n_frames)),
time=2 * np.arange(3 * n_frames))
npt.assert_equal(implant.stim.data.shape,
(implant.n_electrodes, 3 * n_frames))
npt.assert_equal(implant.stim.time, 2 * np.arange(3 * n_frames))
# Verify that a horizontal stimulus will always appear horizontally, even if
# the device is rotated:
data = np.zeros((50, 50))
data[20:-20, 10:-10] = 1
implant.stim = ImageStimulus(data)
model = ScoreboardModel(xrange=(-1, 1),
yrange=(-1, 1),
rho=30,
xystep=0.02)
model.build()
percept = label(model.predict_percept(implant).data.squeeze().T > 0.2)
npt.assert_almost_equal(regionprops(percept)[0].orientation, 0, decimal=1)
def test_ImageStimulus_save():
# Create a horizontal bar:
fname = 'test.png'
shape = (5, 5)
ndarray = np.zeros(shape, dtype=np.uint8)
ndarray[2, :] = 255
imsave(fname, ndarray)
stim = ImageStimulus(fname)
fname2 = 'test2.png'
stim.save(fname2)
npt.assert_almost_equal(stim.data, ImageStimulus(fname2).data)
os.remove(fname)
os.remove(fname2)
def test_ImageStimulus_rotate():
# Create a horizontal bar:
fname = 'test.png'
shape = (5, 5)
ndarray = np.zeros(shape, dtype=np.uint8)
ndarray[2, :] = 255
imsave(fname, ndarray)
stim = ImageStimulus(fname)
# Vertical line:
vert = stim.rotate(90, mode='reflect')
npt.assert_almost_equal(vert.data.reshape(stim.img_shape)[:, 0], 0)
npt.assert_almost_equal(vert.data.reshape(stim.img_shape)[:, 1], 0)
npt.assert_almost_equal(vert.data.reshape(stim.img_shape)[:, 2], 1)
npt.assert_almost_equal(vert.data.reshape(stim.img_shape)[:, 3], 0)
npt.assert_almost_equal(vert.data.reshape(stim.img_shape)[:, 4], 0)
# Diagonal, bottom-left to top-right:
diag = stim.rotate(45, mode='reflect')
npt.assert_almost_equal(diag.data.reshape(stim.img_shape)[0, 4], 1)
npt.assert_almost_equal(diag.data.reshape(stim.img_shape)[2, 2], 1)
npt.assert_almost_equal(diag.data.reshape(stim.img_shape)[4, 0], 1)
npt.assert_almost_equal(diag.data.reshape(stim.img_shape)[0, 0], 0)
npt.assert_almost_equal(diag.data.reshape(stim.img_shape)[4, 4], 0)
# Diagonal, top-left to bottom-right:
diag = stim.rotate(-45, mode='reflect')
npt.assert_almost_equal(diag.data.reshape(stim.img_shape)[0, 0], 1)
npt.assert_almost_equal(diag.data.reshape(stim.img_shape)[2, 2], 1)
npt.assert_almost_equal(diag.data.reshape(stim.img_shape)[4, 4], 1)
npt.assert_almost_equal(diag.data.reshape(stim.img_shape)[0, 4], 0)
npt.assert_almost_equal(diag.data.reshape(stim.img_shape)[4, 0], 0)
os.remove(fname)
def test_ImageStimulus_plot():
# Create a horizontal bar:
fname = 'test.png'
shape = (5, 5)
ndarray = np.zeros(shape, dtype=np.uint8)
ndarray[2, :] = 255
imsave(fname, ndarray)
stim = ImageStimulus(fname)
ax = stim.plot()
npt.assert_equal(ax.axis(), (-0.5, 4.5, 4.5, -0.5))
os.remove(fname)
def test_ImageStimulus_threshold():
# Create a dummy image:
fname = 'test.png'
shape = (25, 37, 3)
gray = 129 / 255.0
create_dummy_img(fname, shape, 'ones', gray=gray)
# Gray levels are between 0 and 1, and can be inverted:
stim = ImageStimulus(fname, as_gray=True)
stim_th = stim.threshold(0.5)
npt.assert_almost_equal(stim.data, gray)
npt.assert_equal(stim.img_shape, shape[:2])
os.remove(fname)
def test_ImageStimulus_center():
# Create a horizontal bar:
fname = 'test.png'
shape = (5, 5)
ndarray = np.zeros(shape, dtype=np.uint8)
ndarray[2, :] = 255
imsave(fname, ndarray)
# Center phosphene:
stim = ImageStimulus(fname)
npt.assert_almost_equal(stim.data, stim.center().data)
npt.assert_almost_equal(stim.data, stim.shift(0, 2).center().data)
os.remove(fname)
def test_ImageStimulus_invert():
# Create a dummy image:
fname = 'test.png'
shape = (25, 37)
gray = 1 / 255.0
create_dummy_img(fname, shape, 'ones', gray=gray)
# Gray levels are between 0 and 1, and can be inverted:
stim = ImageStimulus(fname)
npt.assert_almost_equal(stim.data, gray)
npt.assert_almost_equal(stim.invert().data, 1 - gray)
# Inverting does not change the original object:
npt.assert_almost_equal(stim.data, gray)
os.remove(fname)
def test_ImageStimulus_resize():
fname = 'test.png'
shape = (25, 37, 3)
gray = 129 / 255.0
create_dummy_img(fname, shape, 'ones', gray=gray)
# Gray levels are between 0 and 1, and can be inverted:
stim = ImageStimulus(fname)
npt.assert_almost_equal(stim.data, gray)
npt.assert_equal(stim.resize((13, -1)).img_shape, (13, 19, 3))
# Resize with one dimension -1:
npt.assert_equal(stim.resize((-1, 24)).img_shape, (16, 24, 3))
with pytest.raises(ValueError):
stim.resize((-1, -1))
os.remove(fname)
def test_ImageStimulus_rgb2gray():
# Create a dummy image:
fname = 'test.png'
shape = (25, 37, 3)
gray = 1 / 255.0
create_dummy_img(fname, shape, 'ones', gray=gray)
# Gray levels are between 0 and 1, and can be inverted:
stim_rgb = ImageStimulus(fname)
stim_gray = stim_rgb.rgb2gray()
npt.assert_almost_equal(stim_gray.data, gray)
npt.assert_equal(stim_gray.img_shape, shape[:2])
# Original stim unchanged:
npt.assert_equal(stim_rgb.img_shape, shape)
os.remove(fname)
def test_ImageStimulus():
# Create a dummy image:
fname = 'test.png'
shape = (25, 37, 4)
ndarray = create_dummy_img(fname, shape, 'rand', return_data=True)
# Make sure ImageStimulus loaded is identical to dummy image:
stim = ImageStimulus(fname)
npt.assert_equal(stim.shape, (np.prod(shape), 1))
npt.assert_almost_equal(stim.data, ndarray.reshape((-1, 1)), decimal=2)
npt.assert_equal(stim.metadata['source'], fname)
npt.assert_equal(stim.metadata['source_shape'], shape)
npt.assert_equal(stim.time, None)
npt.assert_equal(stim.electrodes, np.arange(np.prod(shape)))
os.remove(fname)
def test_ImageStimulus_trim():
shape = (13, 29)
ndarray = np.zeros(shape)
ndarray[1:-1, 1:-1] = 0.1
ndarray[2:-2, 2:-2] = 0.2
stim = ImageStimulus(ndarray)
npt.assert_equal(stim.trim().img_shape, (shape[0] - 2, shape[1] - 2))
npt.assert_equal(
stim.trim(tol=0.05).img_shape, (shape[0] - 2, shape[1] - 2))
npt.assert_equal(
stim.trim(tol=0.1).img_shape, (shape[0] - 4, shape[1] - 4))
npt.assert_equal(stim.trim(tol=0.2).img_shape, (1, 0))
npt.assert_equal(
stim.trim(tol=0.1).img_shape,
stim.trim().trim(tol=0.1).img_shape)
def test_ImageStimulus_scale():
# Create a horizontal bar:
fname = 'test.png'
shape = (5, 5)
ndarray = np.zeros(shape, dtype=np.uint8)
ndarray[2, :] = 255
imsave(fname, ndarray)
# Scale phosphene:
stim = ImageStimulus(fname)
npt.assert_almost_equal(stim.data, stim.scale(1).data)
npt.assert_almost_equal(stim.scale(0.1)[12], 1)
npt.assert_almost_equal(stim.scale(0.1)[:12], 0)
npt.assert_almost_equal(stim.scale(0.1)[13:], 0)
with pytest.raises(ValueError):
stim.scale(0)
os.remove(fname)
def test_ImageStimulus_encode():
stim = ImageStimulus(np.random.rand(4, 5))
# Amplitude encoding in default range:
enc = stim.encode()
npt.assert_almost_equal(enc.time[-1], 500)
npt.assert_almost_equal(enc.data.max(axis=1).min(), 0)
npt.assert_almost_equal(enc.data.max(axis=1).max(), 50)
# Amplitude encoding in custom range:
enc = stim.encode(amp_range=(2, 43))
npt.assert_almost_equal(enc.time[-1], 500)
npt.assert_almost_equal(enc.data.max(axis=1).min(), 2)
npt.assert_almost_equal(enc.data.max(axis=1).max(), 43)
with pytest.raises(TypeError):
stim.encode(pulse={'invalid': 1})
with pytest.raises(ValueError):
stim.encode(pulse=LogoUCSB())
def test_ImageStimulus_filter():
# Create a dummy image:
fname = 'test.png'
shape = (25, 37)
create_dummy_img(fname, shape, 'rand')
stim = ImageStimulus(fname)
for filt in ['sobel', 'scharr', 'canny', 'median']:
filt_stim = stim.filter(filt)
npt.assert_equal(filt_stim.shape, stim.shape)
npt.assert_equal(filt_stim.img_shape, stim.img_shape)
npt.assert_equal(filt_stim.electrodes, stim.electrodes)
npt.assert_equal(filt_stim.time, None)
# Invalid filter name:
with pytest.raises(TypeError):
stim.filter({'invalid'})
with pytest.raises(ValueError):
stim.filter('invalid')
os.remove(fname)
def test_ImageStimulus_shift():
# Create a horizontal bar:
fname = 'test.png'
shape = (5, 5)
ndarray = np.zeros(shape, dtype=np.uint8)
ndarray[2, :] = 255
imsave(fname, ndarray)
stim = ImageStimulus(fname)
# Top row:
top = stim.shift(0, -2)
npt.assert_almost_equal(top.data.reshape(stim.img_shape)[0, :], 1)
npt.assert_almost_equal(top.data.reshape(stim.img_shape)[1:, :], 0)
# Bottom row:
bottom = stim.shift(0, 2)
npt.assert_almost_equal(bottom.data.reshape(stim.img_shape)[:4, :], 0)
npt.assert_almost_equal(bottom.data.reshape(stim.img_shape)[4, :], 1)
# Bottom right pixel:
bottom = stim.shift(4, 2)
npt.assert_almost_equal(bottom.data.reshape(stim.img_shape)[4, 4], 1)
npt.assert_almost_equal(bottom.data.reshape(stim.img_shape)[:4, :], 0)
npt.assert_almost_equal(bottom.data.reshape(stim.img_shape)[:, :4], 0)
os.remove(fname)
def test_ImageStimulus_crop():
fname = 'test.png'
shape = (30, 50, 3)
gray = create_dummy_img(fname, shape, 'rand')
stim = ImageStimulus(fname)
stim_cropped = stim.crop(idx_rect=[5, 10, 25, 40])
npt.assert_equal(stim_cropped.img_shape, (20, 30, 3))
npt.assert_equal(
stim_cropped.data.reshape(stim_cropped.img_shape)[3, 7],
stim.data.reshape(stim.img_shape)[8, 17])
npt.assert_equal(
stim_cropped.data.reshape(stim_cropped.img_shape)[10, 28],
stim.data.reshape(stim.img_shape)[15, 38])
npt.assert_equal(
stim.electrodes.reshape(30, 50, 3)[8, 17, 0],
stim_cropped.electrodes.reshape(20, 30, 3)[3, 7, 0])
npt.assert_equal(
stim.electrodes.reshape(30, 50, 3)[15, 38, 2],
stim_cropped.electrodes.reshape(20, 30, 3)[10, 28, 2])
stim_cropped2 = stim.crop(left=10, right=8, top=6, bottom=7)
npt.assert_equal(stim_cropped2.img_shape, (17, 32, 3))
npt.assert_equal(
stim_cropped2.data.reshape(stim_cropped2.img_shape)[3, 7],
stim.data.reshape(stim.img_shape)[9, 17])
npt.assert_equal(
stim_cropped2.data.reshape(stim_cropped2.img_shape)[10, 28],
stim.data.reshape(stim.img_shape)[16, 38])
#"crop-indices and crop-width (left, right, up, down) cannot exist at the same time"
with pytest.raises(ValueError):
stim.crop(idx_rect=[5, 10, 25, 40], left=10)
with pytest.raises(ValueError):
stim.crop([5, 10, 25, 40], right=8)
with pytest.raises(ValueError):
stim.crop([5, 10, 25, 40], top=6)
with pytest.raises(ValueError):
stim.crop([5, 10, 25, 40], bottom=7)
# "crop-width(left, right, up, down) cannot be negative"
with pytest.raises(ValueError):
stim.crop(left=-1)
with pytest.raises(ValueError):
stim.crop(right=-1)
with pytest.raises(ValueError):
stim.crop(top=-1)
with pytest.raises(ValueError):
stim.crop(bottom=-1)
# "crop-width should be smaller than the shape of the image"
with pytest.raises(ValueError):
stim.crop(left=32, right=20)
with pytest.raises(ValueError):
stim.crop(top=12, bottom=18)
# "crop-indices must be on the image"
with pytest.raises(ValueError):
stim.crop([-1, 10, 25, 40])
with pytest.raises(ValueError):
stim.crop([5, -1, 25, 40])
with pytest.raises(ValueError):
stim.crop([5, 10, 31, 40])
with pytest.raises(ValueError):
stim.crop([5, 10, 25, 51])
# "crop-indices is invalid. It should be [y1,x1,y2,x2], where (y1,x1) is upperleft and (y2,x2) is bottom-right"
with pytest.raises(ValueError):
stim.crop([5, 10, 4, 40])
with pytest.raises(ValueError):
stim.crop([5, 10, 25, 9])